Prefix Spreadsheet datafiles disappeared in CP2.1

Hi,

In CP 2.0 the ‘Output Filename’ option for HDF5 and Matlab files was also used as a prefix for the *.CSV files generated in the ExportToSpreadsheet module.
I always used it because it was handy tot distinguish datafiles from different runs. Unfortunately this ‘Output Filename’ is only used for HDF5 and Matlab files.
I can imagine that the previous way it worked was somewhat unclear but very useful. Is it possible to give the the ‘ExportToSpreadsheet’ module an prefix option?

Regards,
Peter

Hi Peter,

Your request is reasonable, though you might want to see our discussion here relating to our desire to distance ourselves from the Default In/Output Folders. But it seems an option to add a default prepend is prudent. Mark added your request here. Mark also has a workaround described there for the near-term.

Cheers,
David

Hi David,

Thank you (and Mark) for your answer. I’ll try the workaround of Mark!

You are also talking about the desire to cancel the Default In/Output Folders. I noticed the smaller role of it in CP 2.1.
For my pipelines these Default In/Output Folders play an important role:

  • The Default input folder was the folder with images to be analysed. In CP 2.1 I have to browse for the images. I prefer the CP 2.0 approach but the new procedure is also workable.
  • I use the Default input folder to define the location of classifiers in ClassifyPixels module (I like to have the classifiers together with the images to be processed).
  • I use the Default output folder to save output images in its sub folder(s) (defined in save images module).
  • I also make new default outputfolders with different version numbers to distinguish the output of different pipelines (analysing the same images).

If the Default In/Output Folders disappear in future does this mean I can’t use the mentioned workflow anymore or will there be alternatives?

Regards,
Peter

Hi Peter,

Here’s a link to our “Roadmap” for our thoughts on UI, etc: github.com/CellProfiler/CellPro … e-1-images

[quote]- The Default input folder was the folder with images to be analysed. In CP 2.1 I have to browse for the images. I prefer the CP 2.0 approach but the new procedure is also workable.
[/quote]

Simply put, our idea is to “Load the images once” then keep a file list. (For really large projects, we prefer to use LoadData at this point to be really clear about image lists and metadata.) When you save a Project file, this also saves the image list, so you don’t need to load it again. For large image lists (e.g. imaging screens) this is a big plus. Note that you can also browse for images while in the Images module: Edit -> Browse For Images. Also, if you import a pre-2.1 pipeline, then the Default Input Folder should show up when you click the “View Output Settings” button. But keep letting us know your feedback!

[quote]- I use the Default input folder to define the location of classifiers in ClassifyPixels module (I like to have the classifiers together with the images to be processed).
[/quote]

In ClassifyPixels’ “Classifier file location” setting, you should be able to see the Default Output (and depending, Input too) Folder there. Or do you mean something else?

[quote]- I use the Default output folder to save output images in its sub folder(s) (defined in save images module).
[/quote]

There will always be an analogue of the “Default Output Folder”. We plan to overhaul the new “Output” section of the pipeline panel, which is now a single button. We mainly wanted to get rid of the DIF/DOF on the main GUI.

This will still be possible – see my previous response.

Cheers,
David

Hi David,

After using CP 2.1 now for some weeks I now learned how to use it and fortunately I can use it like I was used to it in CP 2.0. I can imagine that the new input modules work better for other users. I again now use the ‘old’ LoadImages module instead of the new Input modules and it works fine. I also keep using the DIF/DOF options. You told: “We mainly wanted to get rid of the DIF/DOF on the main GUI.” I will be happy if I can keep using it somehow.

I also found my way to use the ExportToSpreadsheet module and get the files I want but it was confusing. I wanted to get the ‘image.csv’ file but it took some time to get it so far. I chose: ‘Calculate the per-image mean values for object measurements?’ but got no image.csv file. Then I used: ’ Export all measurement types? : No’ and choose the files to export. When I also added ‘Image’ as dataset I got my per-image mean values. So it works but I have to get used to the new GUI and options.

Regards,
Peter

Your preferences are noted! You should be able to choose to not convert your old pipelines and the LoadImages module should function the same as before.

If you add the ExportToSpreadsheet module, it defaults to exporting all the CSVs (Experiment, Image and Objects). If you have a specific suggestion as to how we can improve the settings here, let us know!
David